Energy-saving optimization system for mining explosion-proof frequency converter

Abstract

The invention relates to the technical field of frequency converter energy saving, in particular to a mining flame-proof frequency converter energy-saving optimization system which specifically comprises an upper computer, a switch, a PID controller, a power distribution cabinet, a frequency converter, a driving motor, a belt conveyor, a laser radar sensor, a speed sensor and an intelligent frequency conversion module. In order to solve the problem that delay exists between measurement and control due to the fact that a traditional transportation scale measures the coal flow, the laser radar sensor is used for measuring the coal flow on the belt conveyor in real time through an integral method, and the measurement accuracy is improved while the hysteresis effect is eliminated; in order to solve the problem that a traditional PID control strategy is not high in control precision, the intelligent frequency conversion module is provided, and frequency conversion control over the belt conveyor is achieved in the intelligent frequency conversion module through an energy-saving optimization model based on a BP neural network and a PID control model based on a fuzzy neural network, so that high energy-saving efficiency and high reliability can be kept.

Description

Technical Field

[0001] The invention relates to the technical field of frequency converter energy saving, and in particular to an energy saving optimization system for a mine flameproof frequency converter. Background Art

[0002] With the rapid development of the coal mining industry, the energy consumption problem of mining electrical equipment has become increasingly prominent. The mining explosion-proof frequency conversion regulation system plays a key role in coal mine production. Therefore, in order to improve the operating efficiency of mining electrical equipment, reduce energy consumption, and ensure the safety of coal mine production, the present invention proposes a mining explosion-proof frequency converter energy-saving optimization system, which is mainly used for energy-saving optimization frequency conversion regulation of underground belt conveyors in coal mines. However, the existing mining explosion-proof frequency converter energy-saving optimization system has the following problems:

[0003] (1) The traditional method of measuring coal flow by transport scales results in a time delay between measurement and control, which makes the frequency conversion control response insufficient;

[0004] (2) When using the traditional PID control strategy, the control accuracy is not high under the nonlinear operation of the belt conveyor. Summary of the invention

[0005] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an energy-saving optimization system for explosion-proof frequency converters for mines. In view of the problem that the method of measuring coal flow by traditional transport scales leads to a time delay between measurement and control, making the frequency conversion control response not timely enough, the present invention uses a lidar sensor to obtain the coal laser point cloud data on the belt conveyor, and performs noise reduction processing on it through a K-means clustering algorithm combined with an adaptive threshold filtering algorithm, and then calculates the coal flow in an integral manner, providing data support for frequency conversion regulation, while eliminating the lag effect in traditional measurement methods. The accuracy of measurement is improved, ensuring the accuracy and stability of frequency conversion regulation of belt conveyors in coal mines; the control accuracy is not high when the traditional PID control strategy is used under the nonlinear operation of the belt conveyor. In order to solve the problem of energy conservation and environmental protection, the present invention proposes an intelligent frequency conversion module, and uses an energy-saving optimization model based on a BP neural network and a PID control model based on a fuzzy neural network in the intelligent frequency conversion module to realize automatic energy-saving frequency conversion control of a belt conveyor. The energy-saving optimization model based on the BP neural network is used to learn the relationship among the coal flow rate, the running speed of the conveyor belt and the power of the belt conveyor, and the optimal running speed of the belt conveyor under the condition of fixed coal flow rate is obtained. At the same time, the control strategy is further optimized by using the PID control model based on the fuzzy neural network, and the central value and width of the Gaussian membership function are corrected by using the kinetic energy coefficient, thereby avoiding the problems of low control accuracy and over-adjustment in traditional control strategies, so that the belt conveyor can still maintain high energy-saving efficiency and high reliability when dealing with complex working conditions such as nonlinear and time-varying loads.

[0006] The technical solution adopted by the present invention is as follows: The present invention provides an energy-saving optimization system for a mine explosion-proof frequency converter, which specifically includes: a host computer, a switch, a PID controller, a power distribution cabinet, a frequency converter, a drive motor, a belt conveyor, a laser radar sensor, a speed sensor and an intelligent frequency conversion module;

[0007] The host computer communicates and exchanges data with the PID controller via a switch;

[0008] The switch receives the signal from the host computer and forwards it to the PID controller;

[0009] The PID controller sends a frequency control signal to the frequency converter;

[0010] The power distribution cabinet provides power to the frequency converter;

[0011] The frequency converter receives the control signal from the PID controller and adjusts the frequency to change the speed of the drive motor;

[0012] The driving motor is controlled by a frequency converter to drive the belt conveyor to operate;

[0013] The belt conveyor uses a conveyor belt to transport coal excavated from underground coal mines;

[0014] The laser radar sensor is fixed above the belt conveyor to obtain laser point cloud data of the coal on the belt conveyor;

[0015] The speed sensor is installed on the belt conveyor to obtain the actual running speed of the conveyor belt;

[0016] The intelligent frequency conversion module calculates the coal flow rate according to the laser point cloud data of the coal, and realizes automatic energy-saving frequency conversion control of the belt conveyor by using the energy-saving optimization model based on the BP neural network and the PID control model based on the fuzzy neural network.

[0017] The intelligent frequency conversion module specifically comprises: a coal flow calculation unit, an energy-saving optimization unit and a fuzzy control unit, and the coal flow calculation unit, the energy-saving optimization unit and the fuzzy control unit are unidirectionally connected in sequence.

[0018] The coal flow calculation unit calculates the coal flow through a laser radar sensor, and the specific steps are as follows:

[0019] Step S1: Obtaining laser point cloud data: The laser radar sensor emits laser pulses to the belt conveyor to obtain the laser point cloud data of the coal. The laser point cloud data specifically includes the rotation angle of each laser point and the distance from each laser point to the laser radar sensor;

[0020] Step S2: Establish a Cartesian coordinate system: take the position of the laser radar sensor as the origin, the moving direction of the conveyor belt as the Z axis, the scanning direction of the laser pulse perpendicular to the Z axis with a rotation angle of 0 degrees as the X axis, and the direction perpendicular to the X axis and the Z axis as the Y axis;

[0021] Step S3: constructing a coal cross-section diagram: the plane where the origin, the X axis and the Y axis are located is taken as the OXY plane, and the coal cross-section diagram is composed of the laser point cloud data of the coal in the OXY plane;

[0022] Step S4: using K-means clustering algorithm combined with adaptive threshold filtering algorithm to reduce noise on laser point cloud data;

[0023] Step S5: calculating the area of ​​the coal cross-section;

[0024] Step S6: Calculate the coal flow rate. The specific formula is as follows:

[0025]

[0026] Where Q is the coal flow rate and ρ is the average density of coal, in kg / m 3 , p is the frame rate of the lidar sensor, in FPS, M is the total number of coal cross-section images, v is the running speed of the conveyor belt, in m / s, S total (j) is the area of ​​the jth coal cross section, in m 2 .

[0027] Furthermore, the specific steps of using the K-means clustering algorithm combined with the adaptive threshold filtering algorithm to reduce the noise of the laser point cloud data in step S4 are as follows:

[0028] Step S41: clustering the laser point cloud data using the K-means clustering algorithm;

[0029] Step S42: Calculate the local density of each laser point in the cluster using the Gaussian kernel function. The specific calculation formula is as follows:

[0030]

[0031] Among them, ρ x represents the local density of the laser point x, U represents any cluster, x and x' represent any laser point in U, dis(x,x') represents the Euclidean distance between x and x', and h represents the width parameter of the Gaussian kernel function;

[0032] Step S43: Calculate the standard deviation of the local density of all laser points in the cluster, and use three times the standard deviation as the adaptive threshold of the cluster;

[0033] Step S44: removing laser points in each cluster whose local density is outside the adaptive threshold of the cluster.

[0034] Furthermore, the specific steps of calculating the area of ​​the coal cross-section diagram in step S5 are as follows:

[0035] Step S51: Record two adjacent laser pulses in the laser pulses emitted by the laser radar sensor in the OXY plane to detect coal as a first pulse and a second pulse, wherein the rotation angle of the first pulse is greater than the rotation angle of the second pulse;

[0036] Step S52: The two adjacent laser points generated by the first pulse and the second pulse at the bottom of the coal cross-section are respectively recorded as point A and point B, and the rotation angles of point A and point B are respectively recorded as θ A and θ B ;

[0037] Step S53: two adjacent laser points generated by the first pulse and the second pulse at the top of the coal cross-section are recorded as point C and point D respectively, the rotation angle of point C is consistent with that of point A, and the rotation angle of point D is consistent with that of point B;

[0038] Step S54: Calculate the area of ​​the quadrilateral formed by the four points A, B, C, and D. The specific formula is as follows:

[0039]

[0040] Among them, S ABCD Represents the area of ​​the quadrilateral formed by the four points A, B, C, and D, in m 2 , S ΔOAB represents the area of ​​the triangle formed by the origin, point A and point B, S ΔOCD represents the area of ​​the triangle formed by the origin, point C, and point D, L A , L B , L C and L D Respectively represent the distances from point A, point B, point C, and point D to the lidar sensor, in meters;

[0041] Step S55: The coal cross-section diagram is regarded as a figure composed of a number of quadrilaterals. The calculation formula of the area of ​​the coal cross-section diagram is as follows:

[0042]

[0043] Among them, S total Represents the area of ​​the coal cross section, in m 2 , N represents the total number of laser pulses emitted by the lidar sensor in the OXY plane that detect coal, i is the index of the laser pulse, θ i+1 and θi Respectively represent the rotation angles of the i+1th and ith laser pulses, in degrees, L i+1 , L i Respectively represent the distance from the laser point generated by the i+1th and ith laser pulses at the bottom of the coal cross-section to the lidar sensor, in m, L' i+1 and L' i They respectively represent the distance from the laser point generated by the i+1th and ith laser pulses at the top of the coal cross-section to the lidar sensor, in m.

[0044] The energy-saving optimization unit constructs an energy-saving optimization model based on a BP neural network, learns the relationship between the coal flow rate, the running speed of the conveyor belt and the power of the belt conveyor, and obtains the optimal speed when the coal flow rate is fixed. The optimal speed is the running speed of the conveyor belt when the power of the belt conveyor is minimized. The energy-saving optimization model based on the BP neural network specifically includes a BP input layer, a hidden layer and a BP output layer;

[0045] The BP input layer takes the coal flow rate and the running speed of the conveyor belt as input data;

[0046] The hidden layer learns the input data and uses the Sigmoid activation function;

[0047] The BP output layer learns the output data of the hidden layer and obtains the power of the belt conveyor through a linear activation function.

[0048] The fuzzy control unit uses a PID control model based on a fuzzy neural network to achieve real-time automatic and precise control of the frequency signal of the inverter. The specific steps are as follows:

[0049] Step D1: obtaining the current running speed of the conveyor belt through the speed sensor, and calculating the current coal flow through the coal flow calculation unit;

[0050] Step D2: obtaining the optimal speed corresponding to the current coal flow rate through the energy-saving optimization unit;

[0051] Step D3: The difference between the optimal speed and the current operating speed and the rate of change of the difference are used as input quantities, and the PID control parameters are obtained after calculation through the PID control model based on the fuzzy neural network. The PID control parameters are used to perform frequency conversion control on the frequency converter.

[0052] The PID control model based on fuzzy neural network specifically includes: fuzzy input layer, fuzzification layer, fuzzy rule layer, normalization layer and fuzzy output layer;

[0053] The fuzzy input layer receives input, which includes a difference and a rate of change of the difference;

[0054] The fuzzification layer constructs 7 fuzzy subsets, and fuzzifies each fuzzy subset through a Gaussian membership function;

[0055] The fuzzy rule layer consists of 49 nodes, each node represents a control rule;

[0056] The normalization layer normalizes the output of each node of the fuzzy rule layer;

[0057] The fuzzy output layer outputs PID control parameters;

[0058] The PID control model based on fuzzy neural network sets the first kinetic energy coefficient and the second kinetic energy coefficient, and corrects the center value and the width of the Gaussian membership function. The specific formula is as follows:

[0059]

[0060] in, and represents the central value of the Gaussian membership function of the nth fuzzy subset of the mth input quantity at the tth iteration, t-1th iteration, and t-2th iteration, α represents the first kinetic energy coefficient, β represents the second kinetic energy coefficient, Loss represents the loss value, μ mn represents the central value of the Gaussian membership function of the nth fuzzy subset of the mth input quantity, and denotes the width of the Gaussian membership function of the nth fuzzy subset of the mth input quantity at the tth iteration, t-1th iteration, and t-2th iteration, respectively, mn Represents the width of the Gaussian membership function of the nth fuzzy subset of the mth input quantity.

[0061] The beneficial results achieved by the present invention using the above scheme are as follows:

[0062] (1) In order to solve the problem that the traditional method of measuring coal flow by transport scales leads to a time delay between measurement and control, resulting in insufficient response of frequency conversion control, the present invention uses a lidar sensor to obtain the coal laser point cloud data on the belt conveyor, and performs noise reduction processing on it through a K-means clustering algorithm combined with an adaptive threshold filtering algorithm, and then calculates the coal flow in an integral manner, providing data support for frequency conversion regulation. While eliminating the lag effect in traditional measurement methods, it improves the measurement accuracy and ensures the accuracy and stability of frequency conversion regulation of belt conveyors in coal mines.

[0063] (2) In order to solve the problem of low control accuracy when the belt conveyor is operating nonlinearly when the traditional PID control strategy is adopted, the present invention proposes an intelligent frequency conversion module, and uses the energy-saving optimization model based on the BP neural network and the PID control model based on the fuzzy neural network in the intelligent frequency conversion module to realize automatic energy-saving frequency conversion control of the belt conveyor. The energy-saving optimization model based on the BP neural network is used to learn the relationship among the coal flow rate, the running speed of the conveyor belt and the power of the belt conveyor, and the optimal running speed of the belt conveyor under the condition of fixed coal flow rate is obtained. At the same time, the control strategy is further optimized by using the PID control model based on the fuzzy neural network, and the central value and width of the Gaussian membership function are corrected by using the kinetic energy coefficient, thereby avoiding the problems of low control accuracy and over-adjustment in the traditional control strategy, so that the belt conveyor can still maintain high energy-saving efficiency and high reliability when dealing with complex working conditions such as nonlinear and time-varying loads. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is a structural diagram of an energy-saving optimization system for a mine flameproof frequency converter proposed by the present invention;

[0065] Figure 2 A schematic diagram for constructing a coal cross-section;

[0066] Figure 3 Auxiliary diagram for calculating the area of ​​coal cross-section diagram;

[0067] Figure 4 This is the architecture diagram of the energy-saving optimization model based on BP neural network;

[0068] Figure 5 This is the architecture diagram of the PID control model based on fuzzy neural network.

[0069] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0071] Example 1, see Figure 1 , Figure 1This is a structural diagram of a mine flameproof inverter energy-saving optimization system proposed by the present invention. In this embodiment, a mine flameproof inverter energy-saving optimization system is proposed, and the specific contents are as follows:

[0072] The present invention provides an energy-saving optimization system for a flameproof frequency converter for mining, which specifically comprises: a host computer, a switch, a PID controller, a power distribution cabinet, a frequency converter, a drive motor, a belt conveyor, a laser radar sensor, a speed sensor and an intelligent frequency conversion module;

[0073] The host computer communicates and exchanges data with the PID controller through the switch;

[0074] The switch receives the signal from the host computer and forwards it to the PID controller;

[0075] The PID controller sends a frequency control signal to the frequency converter;

[0076] The power distribution cabinet provides power to the inverter;

[0077] The frequency converter receives the control signal from the PID controller and adjusts the frequency to change the speed of the drive motor;

[0078] The drive motor is controlled by the frequency converter to drive the belt conveyor.

[0079] Belt conveyors use conveyor belts to transport coal excavated from underground coal mines;

[0080] The laser radar sensor is fixed above the belt conveyor to obtain the laser point cloud data of the coal on the belt conveyor;

[0081] The speed sensor is installed on the belt conveyor to obtain the actual running speed of the conveyor belt;

[0082] The intelligent frequency conversion module calculates the coal flow rate according to the laser point cloud data of the coal, and uses the energy-saving optimization model based on the BP neural network and the PID control model based on the fuzzy neural network to realize automatic energy-saving frequency conversion control of the belt conveyor.

[0083] The intelligent frequency conversion module specifically includes: a coal flow calculation unit, an energy-saving optimization unit and a fuzzy control unit, and the coal flow calculation unit, the energy-saving optimization unit and the fuzzy control unit are unidirectionally connected in sequence.

[0084] In this embodiment, the model of the PID controller is Siemens S7-400; the model of the lidar sensor is Pepperl+Fuchs MWC25M-L2M-B16-V15, with a sensing range of 25m; the model of the inverter is Chuangan explosion-proof BPJ2-800 / 1140 mining explosion-proof and intrinsically safe AC inverter, with a rated power of 800KW and a rated voltage of 1140V.

[0085] Example 2, see Figure 2 , Figure 2 In order to construct a schematic diagram of the coal cross-section diagram, this embodiment implements a coal flow calculation unit based on the above embodiment, and the specific contents are as follows:

[0086] The coal flow calculation unit calculates the coal flow through the laser radar sensor. The specific steps are as follows:

[0087] Step S1: Obtaining laser point cloud data: The laser radar sensor emits laser pulses to the belt conveyor to obtain the laser point cloud data of the coal. The laser point cloud data specifically includes the rotation angle of each laser point and the distance from each laser point to the laser radar sensor;

[0088] Step S2: Establish a Cartesian coordinate system: take the position of the laser radar sensor as the origin, the moving direction of the conveyor belt as the Z axis, the scanning direction of the laser pulse perpendicular to the Z axis with a rotation angle of 0 degrees as the X axis, and the direction perpendicular to the X axis and the Z axis as the Y axis;

[0089] Step S3: constructing a coal cross-section diagram: the plane where the origin, the X axis and the Y axis are located is taken as the OXY plane, and the coal cross-section diagram is composed of the laser point cloud data of the coal in the OXY plane;

[0090] Step S4: using K-means clustering algorithm combined with adaptive threshold filtering algorithm to reduce noise on laser point cloud data;

[0091] Step S5: calculating the area of ​​the coal cross-section;

[0092] Step S6: Calculate the coal flow rate. The specific formula is as follows:

[0093]

[0094] Where Q is the coal flow rate and ρ is the average density of coal, in kg / m 3 , p is the frame rate of the lidar sensor, in FPS, M is the total number of coal cross-section images, v is the running speed of the conveyor belt, in m / s, S total (j) is the area of ​​the jth coal cross section, in m 2 .

[0095] Embodiment 3: Based on the above embodiment, this embodiment realizes noise reduction of laser point cloud data. The specific contents are as follows:

[0096] The specific steps of using the K-means clustering algorithm combined with the adaptive threshold filtering algorithm to reduce the noise of the laser point cloud data in step S4 are as follows:

[0097] Step S41: clustering the laser point cloud data using the K-means clustering algorithm;

[0098] Step S42: Calculate the local density of each laser point in the cluster using the Gaussian kernel function. The specific calculation formula is as follows:

[0099]

[0100] Among them, ρ x represents the local density of the laser point x, U represents any cluster, x and x' represent any laser point in U, dis(x,x') represents the Euclidean distance between x and x', and h represents the width parameter of the Gaussian kernel function;

[0101] Step S43: Calculate the standard deviation of the local density of all laser points in the cluster, and use three times the standard deviation as the adaptive threshold of the cluster;

[0102] Step S44: removing laser points in each cluster whose local density is outside the adaptive threshold of the cluster.

[0103] Example 4, see Figure 3 , Figure 3 The auxiliary diagram for calculating the area of ​​the coal cross-section diagram is provided in this embodiment. Based on the above embodiment, the area of ​​the coal cross-section diagram is calculated. The specific contents are as follows:

[0104] The specific steps of calculating the area of ​​the coal cross-section diagram in step S5 are as follows:

[0105] Step S51: Record two adjacent laser pulses in the laser pulses emitted by the laser radar sensor in the OXY plane to detect coal as a first pulse and a second pulse, wherein the rotation angle of the first pulse is greater than the rotation angle of the second pulse;

[0106] Step S52: The two adjacent laser points generated by the first pulse and the second pulse at the bottom of the coal cross-section are respectively recorded as point A and point B, and the rotation angles of point A and point B are respectively recorded as θ A and θ B , the distance between AB is Δθ;

[0107] Step S53: two adjacent laser points generated by the first pulse and the second pulse at the top of the coal cross-section are recorded as point C and point D respectively, the rotation angle of point C is consistent with that of point A, and the rotation angle of point D is consistent with that of point B;

[0108] Step S54: Calculate the area of ​​the quadrilateral formed by the four points A, B, C, and D. The specific formula is as follows:

[0109]

[0110] Among them, S ABCD Represents the area of ​​the quadrilateral formed by the four points A, B, C, and D, in m 2 , S ΔOAB represents the area of ​​the triangle formed by the origin, point A and point B, S ΔOCD represents the area of ​​the triangle formed by the origin, point C, and point D, L A , L B , L C and L D Respectively represent the distances from point A, point B, point C, and point D to the lidar sensor, in meters;

[0111] Step S55: The coal cross-section diagram is regarded as a figure composed of a number of quadrilaterals. The calculation formula of the area of ​​the coal cross-section diagram is as follows:

[0112]

[0113] Among them, S total Represents the area of ​​the coal cross section, in m 2 , N represents the total number of laser pulses emitted by the lidar sensor in the OXY plane that detect coal, i is the index of the laser pulse, θ i+1 and θ i Respectively represent the rotation angles of the i+1th and ith laser pulses, in degrees, L i+1 , L i Respectively represent the distance from the laser point generated by the i+1th and ith laser pulses at the bottom of the coal cross-section to the lidar sensor, in m, L' i+1 and L' i They respectively represent the distance from the laser point generated by the i+1th and ith laser pulses at the top of the coal cross-section to the lidar sensor, in m.

[0114] Example 5, see Figure 4 , Figure 4 This is an architecture diagram of an energy-saving optimization model based on a BP neural network. Based on the above embodiments, this embodiment implements an energy-saving optimization unit. The specific contents are as follows:

[0115] The energy-saving optimization unit constructs an energy-saving optimization model based on the BP neural network, learns the relationship between the coal flow, the running speed of the conveyor belt and the power of the belt conveyor, and obtains the optimal speed when the coal flow is fixed. The optimal speed is the running speed of the conveyor belt when the power of the belt conveyor is the smallest. The energy-saving optimization model based on the BP neural network specifically includes a BP input layer, a hidden layer and a BP output layer;

[0116] The BP input layer takes the coal flow rate and the conveyor belt running speed as input data;

[0117] The hidden layer learns the input data and uses the Sigmoid activation function. The specific formula is as follows:

[0118] Y1 = sigmoid(IW1 + B1);

[0119] Among them, Y1 represents the output of the hidden layer, sigmoid represents the sigmoid activation function, I represents the matrix composed of the running speed of the coal flow conveyor belt, W1 represents the weight parameter of the hidden layer, and B1 represents the bias parameter of the hidden layer.

[0120] The BP output layer learns the output data of the hidden layer and obtains the power of the belt conveyor through a linear activation function. The specific formula is as follows:

[0121] P = Linear(Y1W2+B2);

[0122] Among them, P represents the power of the belt conveyor, Linear represents the linear activation function, W2 represents the weight parameter of the BP output layer, and B2 represents the bias parameter of the BP output layer.

[0123] In this embodiment, the average coal flow rate within five minutes is calculated, and the running speed of the conveyor belt is adjusted until the belt conveyor reaches a stable operation state. The power of the belt conveyor at this time is the optimal belt conveyor power under the average coal flow rate. The average coal flow rate, the running speed of the conveyor belt and the power of the belt conveyor are recorded at this time and used as a set of experimental data. The process is repeated to obtain 79 sets of experimental data, which are divided into training set and test set in a ratio of 8:2, and the energy-saving optimization model based on the BP neural network is trained.

[0124] Example 6, see Figure 5 , Figure 5 This is a structural diagram of a PID control model based on a fuzzy neural network. This embodiment implements a fuzzy control unit based on the above embodiment. The specific contents are as follows:

[0125] The fuzzy control unit uses the PID control model based on the fuzzy neural network to realize real-time automatic and precise control of the frequency signal of the inverter. The specific steps are as follows:

[0126] Step D1: obtaining the current running speed of the conveyor belt through the speed sensor, and calculating the current coal flow through the coal flow calculation unit;

[0127] Step D2: obtaining the optimal speed corresponding to the current coal flow rate through the energy-saving optimization unit;

[0128] Step D3: The difference between the optimal speed and the current operating speed and the rate of change of the difference are used as input quantities, and the PID control parameters are obtained after calculation through the PID control model based on the fuzzy neural network. The PID control parameters are used to perform frequency conversion control on the frequency converter.

[0129] The PID control model based on fuzzy neural network specifically includes: fuzzy input layer, fuzzification layer, fuzzy rule layer, normalization layer and fuzzy output layer;

[0130] The fuzzy input layer receives inputs, which include differences and rates of change of differences;

[0131] The fuzzification layer constructs 7 fuzzy subsets, and fuzzifies each fuzzy subset through the Gaussian membership function. The 7 fuzzy subsets include NB, NM, NS, ZO, PS, PM, and PB, which represent negative large, negative medium, negative small, zero, positive small, positive medium, and positive large, respectively. The specific calculation formula of the fuzzification layer is as follows:

[0132]

[0133] Among them, Out2(m,n) t represents the output of the nth fuzzy subset of the mth input in the fuzzification layer in the tth iteration, m represents the index of the input, when m=1, it is the difference, when m=2, it is the rate of change of the difference, n represents the index of the fuzzy subset, X mn represents the output of the fuzzy input layer, represents the central value of the Gaussian membership function of the nth fuzzy subset of the mth input quantity in the tth iteration, Represents the width of the Gaussian membership function of the nth fuzzy subset of the mth input quantity in the tth iteration.

[0134] The fuzzy rule layer consists of 49 nodes, each node represents a control rule, and the specific calculation formula is as follows:

[0135] Out3(k) t =Out2(1,n1) t ×Out2(2,n2) t ;

[0136] Among them, Out3(k) t represents the output of the kth node of the fuzzy rule layer in the tth iteration, Out2(1,n1) t The fuzzification layer output of the n1th fuzzy subset of the difference in the tth iteration, Out2(2,n2) t The fuzzification layer output of the n2th fuzzy subset representing the rate of change of the difference in the tth iteration.

[0137] The normalization layer normalizes the output of each node in the fuzzy rule layer. The specific calculation formula is as follows:

[0138]

[0139] Among them, Out4(k) t represents the output of the kth node of the normalization layer in the tth iteration, Out3(a) t Represents the output of the ath node in the fuzzy rule layer in the tth iteration.

[0140] The fuzzy output layer outputs the PID control parameters, which include K p , K i , K d , the specific calculation formula is as follows:

[0141]

[0142] Among them, ω t represents the weight of the tth iteration.

[0143] The calculation formula of the loss function of the PID control model based on fuzzy neural network is as follows:

[0144]

[0145] Among them, Loss t Represents the loss value of the tth iteration, rin t represents the ideal value of the tth iteration, Represents the output result of the fuzzy output layer of the tth iteration.

[0146] The first kinetic energy coefficient and the second kinetic energy coefficient are set to correct the center value and the width of the Gaussian membership function. The specific formula is as follows:

[0147]

[0148] in, and represents the central value of the Gaussian membership function of the nth fuzzy subset of the mth input quantity at the t-1th iteration and the t-2th iteration, α represents the first kinetic energy coefficient, β represents the second kinetic energy coefficient, Loss represents the loss value, μ mn represents the central value of the Gaussian membership function of the nth fuzzy subset of the mth input quantity, and denotes the width of the Gaussian membership function of the nth fuzzy subset of the mth input quantity at the t-1th iteration and the t-2th iteration, σ mnRepresents the width of the Gaussian membership function of the nth fuzzy subset of the mth input quantity.

Claims

1. A mining flameproof inverter energy-saving optimization system, characterized by: Specifically include: A host computer, a switch, a PID controller, a power distribution cabinet, a frequency converter, a drive motor, a belt conveyor, a laser radar sensor, a speed sensor and an intelligent frequency conversion module, wherein the belt conveyor comprises a conveyor belt, and the conveyor belt transports coal; The intelligent frequency conversion module specifically includes: a coal flow calculation unit, an energy-saving optimization unit and a fuzzy control unit, and the coal flow calculation unit, the energy-saving optimization unit and the fuzzy control unit are connected in one direction in sequence; The coal flow calculation unit calculates the coal flow through a laser radar sensor, and the specific steps are as follows: Step S1: Obtaining laser point cloud data: The laser radar sensor emits laser pulses to the belt conveyor to obtain the laser point cloud data of the coal. The laser point cloud data specifically includes the rotation angle of each laser point and the distance from each laser point to the laser radar sensor; Step S2: Establish a Cartesian coordinate system: take the position of the laser radar sensor as the origin, the moving direction of the conveyor belt as the Z axis, the scanning direction of the laser pulse perpendicular to the Z axis with a rotation angle of 0 degrees as the X axis, and the direction perpendicular to the X axis and the Z axis as the Y axis; Step S3: constructing a coal cross-section diagram: the plane where the origin, the X axis and the Y axis are located is taken as the OXY plane, and the coal cross-section diagram is composed of the laser point cloud data of the coal in the OXY plane; Step S4: using K-means clustering algorithm combined with adaptive threshold filtering algorithm to reduce noise on laser point cloud data; Step S5: calculating the area of ​​the coal cross-section; Step S6: Calculate the coal flow rate. The specific formula is as follows: Where Q is the coal flow rate, ρ is the average density of coal, in kg / m 3 , p is the frame rate of the lidar sensor, in FPS, M is the total number of coal cross-section images, v is the running speed of the conveyor belt, in m / s, S total (j) is the area of ​​the jth coal cross section, in m 2 .

2. The energy-saving optimization system for a mine flameproof frequency converter according to claim 1 is characterized in that: The specific steps of using the K-means clustering algorithm combined with the adaptive threshold filtering algorithm to reduce the noise of the laser point cloud data in step S4 are as follows: Step S41: clustering the laser point cloud data using the K-means clustering algorithm; Step S42: Calculate the local density of each laser point in the cluster using the Gaussian kernel function. The specific calculation formula is as follows: Among them, ρ x represents the local density of the laser point x, U represents any cluster, x and x' represent any laser point in U, dis(x,x') represents the Euclidean distance between x and x', and h represents the width parameter of the Gaussian kernel function; Step S43: Calculate the standard deviation of the local density of all laser points in the cluster, and use three times the standard deviation as the adaptive threshold of the cluster; Step S44: removing laser points in each cluster whose local density is outside the adaptive threshold of the cluster.

3. The energy-saving optimization system for a mine flameproof frequency converter according to claim 2 is characterized in that: The energy-saving optimization unit constructs an energy-saving optimization model based on a BP neural network to obtain the optimal speed under a fixed coal flow rate, which is the running speed of the conveyor belt when the power of the belt conveyor is minimized.

4. The energy-saving optimization system for a mine flameproof frequency converter according to claim 2 is characterized in that: The fuzzy control unit uses a PID control model based on a fuzzy neural network to achieve real-time automatic and precise control of the frequency signal of the inverter. The specific steps are as follows: Step D1: obtaining the current running speed of the conveyor belt through the speed sensor, and calculating the current coal flow through the coal flow calculation unit; Step D2: obtaining the optimal speed corresponding to the current coal flow rate through the energy-saving optimization unit; Step D3: The difference between the optimal speed and the current operating speed and the rate of change of the difference are used as input quantities, and the PID control parameters are obtained after calculation through the PID control model based on the fuzzy neural network. The PID control parameters are used to perform frequency conversion control on the frequency converter.

5. The energy-saving optimization system for a mine flameproof frequency converter according to claim 4 is characterized in that: The PID control model based on fuzzy neural network specifically includes: a fuzzy input layer, a fuzzification layer, a fuzzy rule layer, a normalization layer and a fuzzy output layer.

6. The energy-saving optimization system for a mine flameproof frequency converter according to claim 5 is characterized in that: The fuzzification layer constructs seven fuzzy subsets, and fuzzifies each fuzzy subset through a Gaussian membership function.

7. The energy-saving optimization system for a mine flameproof frequency converter according to claim 6 is characterized in that: The PID control model based on fuzzy neural network sets the first kinetic energy coefficient and the second kinetic energy coefficient, and corrects the center value and the width of the Gaussian membership function. The specific formula is as follows: in, and represents the central value of the Gaussian membership function of the nth fuzzy subset of the mth input quantity at the tth iteration, t-1th iteration, and t-2th iteration, α represents the first kinetic energy coefficient, β represents the second kinetic energy coefficient, Loss represents the loss value, μ mn represents the central value of the Gaussian membership function of the nth fuzzy subset of the mth input quantity, and denotes the width of the Gaussian membership function of the nth fuzzy subset of the mth input quantity at the tth iteration, t-1th iteration, and t-2th iteration, respectively, mn Represents the width of the Gaussian membership function of the nth fuzzy subset of the mth input quantity.

Images